Collective strong light-matter coupling in hierarchical microcavity

Nov 30, 2018 - Ankit Bisht , Jorge Cuadra , Martin Wersäll , Adriana Canales , Tomasz J. Antosiewicz , and Timur Shegai. Nano Lett. , Just Accepted ...
0 downloads 0 Views 4MB Size
Subscriber access provided by Gothenburg University Library

Communication

Collective strong light-matter coupling in hierarchical microcavity-plasmon-exciton systems Ankit Bisht, Jorge Cuadra, Martin Wersäll, Adriana Canales, Tomasz J. Antosiewicz, and Timur Shegai Nano Lett., Just Accepted Manuscript • DOI: 10.1021/acs.nanolett.8b03639 • Publication Date (Web): 30 Nov 2018 Downloaded from http://pubs.acs.org on December 1, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 24 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Collective strong light-matter coupling in hierarchical microcavityplasmon-exciton systems Ankit Bisht1, Jorge Cuadra1, Martin Wersäll1, Adriana Canales1, Tomasz J. Antosiewicz1,2, Timur Shegai1* 1Department 2Faculty

of Physics, Chalmers University of Technology, 412 96, Göteborg, Sweden

of Physics, University of Warsaw, Pasteura 5, 02-093 Warsaw, Poland

* Email: [email protected] Abstract: Polaritons are compositional light-matter quasiparticles that arise as a result of strong coupling between the vacuum field of a resonant optical cavity and electronic excitations in quantum emitters. Reaching such a regime is often hard, as it requires materials possessing high oscillator strengths to interact with the relevant optical mode. Two-dimensional transition metal dichalcogenides (TMDCs) have recently emerged as promising candidates for realization of strong coupling regime at room temperature. However, these materials typically provide coupling strengths in the range of 10-40 meV, which may be insufficient for reaching strong coupling with low quality factor resonators. Here, we demonstrate a universal scheme that allows a straightforward realization of strong coupling with 2D materials and beyond. By intermixing plasmonic excitations in nanoparticle arrays with excitons in a WS2 monolayer inside a resonant metallic microcavity, we fabricate a hierarchical system with the collective microcavity-plasmon-exciton Rabi splitting exceeding 500 meV at room temperature. Photoluminescence measurements of the coupled systems show dominant emission from the lower polariton branch, indicating the participation of excitons in the coupling process. Strong coupling has been recently suggested to affect numerous optical- and material-related properties including chemical reactivity, exciton transport and optical nonlinearities. With the universal scheme presented here, strong coupling across a wide spectral range is within easy reach and therefore exploring these exciting phenomena can be further pursued in a much broader class of materials. Keywords: Strong plasmon-exciton coupling, TMDC, monolayer WS2, collective Rabi splitting

1 ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 24

Strong coupling between the vacuum field of a cavity and a quantum emitter can be achieved when the rate of energy exchange between the emitter and the cavity becomes faster than the decoherence rate of both the emitter and the cavity. This regime manifests itself in formation of light-matter hybrid states - polaritons

1, 2.

In contrast, weak

coupling between a cavity and an emitter does not result in formation of polaritons, but instead leads to the modification of the spontaneous decay rate, known as the Purcell effect 3.

Strong coupling has been previously demonstrated in a plethora of cavity-emitter

configurations. Traditionally, high quality factor cavities, such as those based on distributed Bragg reflectors (DBR), whispering gallery modes and photonic crystals, are employed for polaritonic applications 1. Major drawbacks associated with the use of such cavities are fabrication challenges and the need for cryogenic temperatures. An alternative methodology of utilizing plasmonic nanoparticles for strong coupling applications offers an advantage of smaller mode volumes and open cavity configurations, which in turn allows for room temperature operation 2, 4. On the excitonic side of the problem, semiconductor quantum wells were traditionally employed for the purposes of strong coupling 1. However, small exciton binding energies of conventional semiconductor materials, such as GaAs, prohibit their use at room temperature 5. To warrant room temperature operation, organic chromophores and Jaggregates are used ubiquitously for strong coupling purposes. However, they suffer from inhomogeneous broadening, disorder, and photostability issues, thereby hindering their use in practical applications 6-8. Recently, monolayer transition metal dichalcogenide (TMDC) semiconductors have emerged as promising alternatives to both semiconductor quantum wells and organic chromophores for exploring polariton physics at room temperature. Monolayer TMDCs possess a direct band gap transition, large exciton binding energies, absorption exceeding 15% at the A-exciton resonance and narrow line width even at room temperature

9-11.

These properties make monolayer TMDCs extremely promising for studying polaritonrelated phenomena. Indeed, plasmonic modes in single nanoparticles have been hybridized with excitons in WS2 and WSe2 in various configurations

12-15

with typical Rabi splitting

reaching 80-120 meV. Similarly nanoparticle lattice modes coupled to MoS2 monolayers have resulted in 100 meV splitting

16, 17.

Strong coupling of both DBR-based and metallic

2 ACS Paragon Plus Environment

Page 3 of 24 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

microcavities with monolayer TMDCs were shown to produce similar Rabi splitting values of 20-100 meV 18-20. More generally plasmon-exciton interactions require strong intermixing between optical and electronic excitations. Such mixing often results in regimes which can be classified as weak, or just at the border between weak and strong 21. However, to explore rich polaritonic physics, an unambiguous observation of strong or even ultra-strong coupling is required. Typically, this is achieved using solid state architectures microcavities

6-8

or lattice resonances

23-25.

1, 5, 22,

organic

However, these approaches suffer from a

necessity to saturate the cavity with a large number of excitonic material(s). Alternatively, strong coupling can be reached by utilizing excitons with high transition dipole moments. However, this reduces the application of strong coupling to a limited class of materials possessing sufficiently high oscillator strengths, such as J-aggregates, quantum dots, perovskites, TMDCs, etc. 4 Here, we present an alternative path to circumvent the problem of reaching strong coupling in nanophotonic systems. The central idea of this study is to realize the strong coupling regime by intermixing plasmons, excitons and microcavity photons into a common polaritonic state. The oscillator strength of plasmonic nanoparticles is orders of magnitude higher than that of TMDCs and molecular excitons, which allows the latter to “borrow” the oscillator strength from the former in order to share the same coherent polaritonic states. The microcavities used in our study consist of two 40 nm thick Au mirrors separated by a dielectric spacer of variable thickness, which are filled with plasmonic nanoparticle arrays and monolayers of WS2. By incorporating the system components in this way, we substantially increase the combined cavity-plasmon-exciton interaction strength. The energy-momentum dispersion for polaritonic states shows massive collective microcavity-plasmon-exciton Rabi splitting exceeding 500 meV – far into the strong coupling regime. These values are enhanced in comparison to any of the twocomponent systems - cavity and monolayer WS2 or Au nanoparticles and monolayer WS2. Photoluminescence measurements of the coupled systems show dominant emission from the lower polariton branch, thereby suggesting strong intermixing of excitons with plasmons and cavity photons in our nanophotonic structures. These results provide a universal recipe to reach the strong coupling regime of interaction and pave the way towards exploring its new and emerging applications. 3 ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 24

The concept. Let us consider a resonant microcavity loaded with a large number of absorbing molecules (see Fig. 1). In the many-emitter strong coupling picture, a large number of emitters coherently contribute to the coupling process to produce the collective Rabi splitting that scales as the square root of the number of involved molecules 𝑁 26. Because the process is coherent, the molecular contribution can be thought of as a giant harmonic oscillator with the oscillator strength equal to that of 𝑁 coherently combined molecules. This in turn implies that all molecules in the coupled system can be replaced by a single entity possessing correspondingly higher oscillator strength, without significantly affecting the mode picture. This concept is illustrated schematically in Fig. 1 (see Methods for further details).

Figure 1. Schematic demonstration of electromagnetic similarity between (a) a cavity coupled to a large number of randomly oriented molecular emitters and (b) a cavity coupled to large dipole moment plasmonic nanoparticles with fewer emitters. (c, d) Corresponding calculated reflection spectra show similar Rabi splittings of 280 meV for both cavity configurations. These observations demonstrate the principle idea of this work, which allows reaching the collective strong coupling regime with fewer emitters.

Such an approach is useful for the following reasons. Consider the transition dipole moment of typical excitons. These range in between a few Debyes for fluorophore molecules and few tens of Debyes for quantum dots, perovskites and 2D materials, 𝜇𝑒=150 Debye 4. Reaching the regime of strong coupling with these excitons requires dense excitonic packing

27

or high-Q microcavity constraints. Contrary to that, plasmonic 4 ACS Paragon Plus Environment

Page 5 of 24 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

nanoantennas can easily support transition dipole moments of several thousands of Debyes, owing to the collective electronic nature of plasmonic excitations. In other words, since the absorption cross-section of a typical dye molecule at room temperature is about 10-16 cm2, while that of a typical plasmonic nanoparticle is 10-10 cm2, a single plasmonic nanoparticle acts as efficiently as approximately 106 coherently excited organic molecules in the collective strong coupling picture. Thus, by intermixing plasmons and excitons within a common microcavity, it is possible to realize collective microcavity-plasmon-exciton polaritons with fewer emitters than in the pure microcavity-exciton case (see Fig. 1). Conceptually our approach is similar to organic-inorganic polaritons pioneered by Agranovich and co-workers

28

and to donor-acceptor polaritons realized experimentally

by Barnes group 29 and followed by a number of more recent publications 30-33. Additionally, microcavities coupled to Au nanospheres were studied previously in the weak-coupling regime, where red-shift of cavity mode was observed due to the presence of the plasmonic particle

34.

Strong coupling of a cavity and a single plasmonic nanorod was recently

demonstrated using photoluminescence spectroscopy

35.

Coupling between microcavities

and plasmonic arrays at infrared frequencies has also been previously shown 36. In addition, plexciton nanoparticles were shown to hybridize with propagating surface plasmon polaritons of a thin Ag film 37. Finally, hybrid systems consisting of plasmonic particles and quantum emitters embedded in microcavities have been studied theoretically 38, 39. Here, we realize similar concepts experimentally by incorporating monolayer WS2 and plasmonic nanoparticle arrays into a common microcavity system. Nanoparticle plasmon–microcavity photon polaritons. In order to study the coupling within the three-component system, we first realized the coupling between a metallic Fabry-Pérot (FP) microcavity coupled to periodic arrays of Au nanodisks (see Fig. 2a). The thickness of both top and bottom mirrors was set to 40 nm, while the thickness of the cavity was varied from 160 nm to 200 nm. Plasmonic arrays of Au disks (height=20 nm, diameter=80 nm) arranged in square arrays with different particle-to-particle distances (Λ =200 nm, 250 nm and 300 nm) were fabricated inside the microcavities. The array spacings were chosen such that the dispersive surface lattice modes are in the blue or UV region – far away from the localized plasmon resonances of individual disks.

5 ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 24

Figure 2. Nanoparticle plasmon-photon polaritons in a microcavity system. (a) Schematic of a microcavity coupled to a square array of Au nanodisks. Here L represents the cavity thickness, d is the nanoparticle diameter and Λ is the array spacing. (b) SEM image of rectangular array of Au nanodisks with diameter d = 80 nm, height h = 20 nm and Λ = 200 nm. The scale bar is 400 nm. (c) Coupling strength g as a function of Λ 1/2 1/2 extracted from experimental data using 𝑔 ≈ (𝜔 + ― 𝜔𝑝𝑙) (𝜔𝑝𝑙 ― 𝜔 ― ) , where 𝜔 + , 𝜔 ― and 𝜔𝑝𝑙 are

energies of upper polariton, lower polariton and nanoparticle localized plasmon. Dashed lines show inverse scaling (g  Λ-1) for all cavity thicknesses. (d-f) Reflection measurements showing microcavity detuning with respect to the nanoparticle plasmon resonance for three different cavity thicknesses: 160 nm (d), 180 nm (e) and 200 nm (f), with each microcavity containing arrays with d = 80 nm, h = 20 nm and Λ = 200 nm, 250 nm and 300 nm. (g-i) FDTD calculations corresponding to experimental reflection in (d-f).

Experimental reflection spectra of bare cavities and cavities loaded with different plasmonic arrays as a function of cavity thickness are shown in Fig. 2d-f. Changing the thickness of the cavity allows controlling the detuning between the cavity and plasmon resonances. The bare cavity mode splits into two distinct dips in the reflection spectra upon interaction with plasmonic arrays, signaling the formation of polaritonic states. Here, these

6 ACS Paragon Plus Environment

Page 7 of 24 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

hybrid states are a mixture of the fundamental cavity mode and localized surface plasmon resonances in Au nanoparticles. These states thus have no excitonic contribution at all. We further compare experimental findings in Fig. 2d-f with numerical calculations, which were performed using the Finite-Difference Time-Domain (FDTD) method. Exact parameters of the coupled systems extracted from scanning electron microscopy (SEM) images were used in the calculations (example of SEM image is shown in Fig. 2b). The agreement between the calculated (Fig. 2g-i) and experimental (Fig. 2d-f) spectra is remarkable. In this case we find that the plasmon-cavity system is comfortably in the strong coupling regime, since the Rabi splitting significantly exceeds both plasmon and microcavity line widths Ω ≫ 𝛾𝑝𝑙 , 𝛾𝑐𝑎𝑣 where 𝛾𝑝𝑙  200 meV, 𝛾𝑐𝑎𝑣  150 meV and Ω  480 meV with the splitting to damping ratio as high as Ω 𝛾𝑐𝑎𝑣  3.2. The plasmon line width was obtained by collecting dark field scattering spectra from Au NP arrays fabricated on glass substrates as shown in Supporting Information (SI, Fig. S1). The cavity line width changes from ~136 meV to 86 meV as the cavity thickness is increased from 160 nm to 200 nm. To account for all possible values of the cavity line width and draw definitive conclusions about the interaction regime, we therefore take an overestimated value of 𝛾𝑐𝑎𝑣  150 meV for the cavity line width in all subsequent calculations. Since the coupling strength is given by 𝑔 = 𝑁𝜇𝑝𝑙|𝐸vac|, where 𝑁 is controlled by the density of Au nanodisks in the arrays, the transition dipole moment of individual Au disks, 𝜇𝑝𝑙, can be experimentally determined from the measured Rabi splitting values. We observe the

𝑁 behavior of mode splitting with the nanoparticle density, similar to the

situation of 𝑁 emitters coherently coupled to a cavity (Fig. 2c). Using the standard expression for the coupling strength, in this case we obtain: 𝑔 ≈ 𝜇𝑝𝑙

ℏ𝜔

,

(1)

2𝜀𝜀0𝐿Λ2

where Λ is the array spacing and 𝐿 is the cavity thickness. Notice that g  Λ-1 in this case. For experimentally measured values 𝑔  280 meV, 𝜀 = 2.25, Λ = 200 nm and 𝐿 = 180 nm, we estimate 𝜇𝑝𝑙 to be about 1.3x104 Debyes. This high value is nearly three orders of magnitude higher than a transition dipole moment of A-exciton in monolayer WS2 and any known molecular or quantum dot transition4. This is precisely what makes plasmonic nanostructures so useful for reaching the collective strong coupling regime as we will demonstrate further. 7 ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 24

Plasmon–exciton–microcavity photon polaritons. We proceed to explore the composition of the polaritonic mixtures by adding the WS2 monolayer to the system. The plasmon, exciton, and microcavity coupled systems were fabricated by incorporating large area mechanically exfoliated monolayer WS2 flakes inside the microcavities and depositing arrays of Au NPs (see Methods). We intentionally fabricated arrays of plasmonic nanoparticles such that they do not cover the whole area of the flake. This allows us to directly compare different polariton compositions using the same flake, as shown in Fig. S2. The WS2 monolayers were exfoliated from bulk crystals and transferred inside half microcavities using PDMS stamps (dry transfer technique)

40.

The WS2 monolayer was

characterized with optical contrast and Raman spectroscopy (see Fig. S3). We first compare the two-component (cavity and WS2) versus the threecomponent

(cavity,

nanoparticle

and

WS2)

samples

using

reflection

(R)

and

photoluminescence (PL) spectroscopy (Fig. 3). The experiments were performed for a range of cavity thicknesses (140 nm - 180 nm) to tune the cavity with respect to both the plasmon and exciton resonances. The coupled systems clearly show hybridized reflection dips in both two-component (Fig. 3a) and three-component (Fig. 3d) systems. However, the mode splitting in the latter case is significantly higher, owed to the plasmonic nanostructures. FDTD calculations (Fig. 3b,e) performed for these samples agree well with the experimental results. The dielectric function for the monolayer WS2 in this case was extracted from Li et al.11 We note that for the three-component system, calculations (Fig. 3e) indicate emergence of middle polariton (MP) states, which are nearly absent in experiments (Fig. 3d). We attribute this to differences in polariton line width between experiments and calculations (in experiments it is broader as is evidenced in Fig. 2 and Fig. 3), which may appear as a result of underestimated losses in the Johnson and Christy permittivity of gold41 that was used in the calculations as well as slight inhomogeneous broadening of nanoparticle plasmons. We also note that control experiments on Au nanodisks deposited on WS2 monolayer outside the cavity do not result in the strong coupling regime as the observed splitting is small in comparison to the nanoparticle plasmon line width (see Fig. S4). We further investigate the strongly coupled samples using PL spectroscopy. For the case of two-component systems, we find that the PL spectra strongly depend on the cavity thickness (Fig. 3c). This is expected as the cavity thickness determines the cavityexciton detuning, and thus composition of polaritonic states in this case. For the case of the 8 ACS Paragon Plus Environment

Page 9 of 24 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

thinnest cavity (140 nm, black curve in Fig. 3c), PL emission occurs primarily at the A-exciton band of WS2 at around 2 eV, indicating that the polaritonic mixture is mostly of excitonic character (see Hopfield coefficients in Fig. S5a). As the cavity thickness is increased, emission via the lower polariton (LP) branch shifts to lower energy 1.76-1.9 eV depending on the exact cavity thickness. This demonstrates stronger involvement of the cavity mode in the formation of the LP state, consistent with the red-shift of cavity resonance for thicker cavities. Noteworthy, PL emission maxima correspond to the LP maxima measured in reflection (see Fig. 3a-c), in agreement with PL emission from strongly coupled molecular 42-44

7,

and TMDC 20 systems. For the three-component systems, in contrast to two-component counterparts,

we find that PL emission is less dependent on the cavity thickness. This can be understood by appreciating the fact that the combined collective Rabi splittings in the three-component cases are significantly stronger (up to 500 meV) than in two-component systems, thereby implying that detuning between the cavity mode and the plasmon and exciton resonances should play a less important role. Here, we again observe that the PL signal is dominated by emission from LP states, which occur at 1.6-1.7 eV depending on the cavity thickness. Noteworthy, the LP states in the three-component system lie significantly below the corresponding two-component counterparts, because of the stronger combined Rabi splitting in this case. We also note that PL measurements were performed on WS2 flakes at intermediate steps during the fabrication process and no additional spectral features were observed (see Fig. S6). We thus conclude that our fabrication procedure does not significantly degrade the quality of WS2 monolayer and that it is relatively free from defect/bound states in the flakes when observed at our experimental conditions. This eliminates the possibility of the low energy PL emission arising from the defect states filtered through the cavity modes. It is further worth noting that the background PL signal arising from bare Au NPs in FP cavity is much lower than the PL of the three-component system. Specifically, the PL of the three-component system at the LP position shows nearly no background signal originating from Au NPs (see Fig. S7).

9 ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 24

Figure 3. Reflectivity (R) and photoluminescence (PL) spectra for samples with different cavity thicknesses (140 nm - 180 nm). (a, d) Experimental near-normal incidence reflection spectra for two-component (a) and three-component (d) systems, correspondingly. (b, e) Calculated reflectivity spectra for two-component (b) and three-component (e) systems, correspondingly. (c, f) Experimental photoluminescence spectra for twocomponent (c) and three-component (f) systems, correspondingly for excitation wavelength of 532 nm (2.33 eV). Note that PL emission follows the lower polariton branch.

Polariton dispersion. Dispersion of polaritonic modes is an essential characteristic of strongly coupled microcavities. In Fig. 4 we show dispersion of 140, 160 and 180 nm samples as a function of polaritonic composition measured using TE-polarized incidence. Dispersion in reflection was obtained by spectrally scanning the back focal plane of the microscope objective with the liquid crystal filter (LCF, see Methods). The data includes (i) empty cavities, (ii) cavities loaded with monolayer of WS2, (iii) cavities loaded 10 ACS Paragon Plus Environment

Page 11 of 24 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

with plasmonic nanoparticles and (iv) cavities loaded with both monolayer WS2 and plasmonic nanoparticles. For the cases when polaritonic mixtures have an excitonic component we complement the reflectivity data with dispersion in PL. As the cavity thickness is increased, we observe a gradual red shift of the bare cavity mode resonance (Fig. 4a-c), in agreement with the standard microcavity behavior. By introducing the WS2 monolayer into the cavity, the exciton hybridizes with the cavity resulting in UP and LP states, whose exact composition and dispersion depend on the cavity thickness and parameters of the 2D material. PL of these two-component samples show parabolic dispersion and dominant emission via the LP state (Fig. 4d-f), in agreement with previous findings18, 20 and normal incidence data in Fig. 3a-c. Dispersion in the microcavity WS2 monolayer system shows mode anti-crossing at 30⁰ (for 160 nm cavity thickness, see Fig. 4e and SI Fig. S5a). The cavity-exciton hybridization results in Rabi splitting of 75 meV. The obtained splitting is at the border between weak and strong coupling regime, since the cavity and exciton line width are 𝛾𝐹𝑃  150 meV and 𝛾𝑋  30 meV correspondingly

20,

resulting in 𝛺 ≲ (𝛾𝑋 + 𝛾𝐹𝑃) 2 = 90 meV. Cavities loaded with plasmonic arrays end up deep in the strong coupling regime, in contrast to two-component cavity-WS2 samples. Dispersion curves in this case show anti-crossing with a substantial pure photonic Rabi-like splitting exceeding 400 meV (Fig. 4g-i). For the 180 nm sample at normal incidence the LP is observed at 1.7 eV, while the UP is observed at 2.2 eV although less pronounced. The mode splittings in this case are slightly smaller than the ones obtained in Fig. 2, because the Au disks diameter in this case was slightly smaller than 80 nm. When the microcavity, monolayer WS2 and Au nanodisks are combined together, the LP is observed at 1.57 eV (at normal incidence), while the UP appears at 2.15 eV resulting in a massive collective microcavity-plasmon-exciton Rabi splitting of about 535 meV (see Fig. 4l). Thus, the mode splitting obtained in the three-component system is significantly larger than any of the two-component systems: cavity-WS2 (75 meV) and cavity-nanoparticles (400 meV) and even larger than their combination. The latter strongly suggests existence of an additional mechanism of the coupling strength enhancement in this case, which is likely due to the antenna effect as we argue below.

11 ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 24

Figure 4. Dispersion in reflection and photoluminescence of several microcavity samples as a function of cavity thickness. (a-c) Dispersion for empty cavities, (d-f) Dispersion for cavities loaded with monolayer WS2, (g-i) Dispersion for cavities loaded with plasmonic nanoparticles and (j-l) Dispersion for cavities loaded with monolayer WS2 and plasmonic nanoparticles. Dashed green lines represent UP and LP states extracted from the coupled oscillator model. White horizontal lines in (j-l) represent the position of WS2 exciton.

To improve our understanding of the coupled system, we investigated its PL response. We used a 532 nm (2.33 eV) continuous wave laser to resonantly excite the UP branch. Similarly to reflection measurements, dispersion in PL was obtained by scanning the back focal plane of the microscope objective with the LCF. PL spectra of three samples of different cavity thicknesses are shown in Fig. 4. For 140 nm thickness cavity, the PL signal is heavily dominated by the LP branch (Fig. 4j). We note that the resonance position of the LP branch is significantly red-shifted with respect to the resonance of uncoupled WS2 12 ACS Paragon Plus Environment

Page 13 of 24 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

monolayer. Such behavior can be explained by a phonon-assisted population of LP via exciton reservoir in the strong coupling regime 45, 46. The PL data thus points to a significant excitonic contribution to formation of polaritons in the three-component system. This observation is further corroborated by the coupled oscillator model and Hopfield coefficients (see Fig. S5) and by FDTD calculations of absorption spectra of the threecomponent system (see Figs. S9-10). As the cavity thickness is increased to 160 nm and 180 nm, the LP emission contribution gets slightly red-shifted as the cavity contribution to LP gradually increases. This is evidenced from dispersion plots in Fig. 4j-l and near-normal incidence spectra in Fig. 3f. For the 180 nm sample, the sample with the most red-detuned cavity mode, PL emission occurs at a slightly higher energy than the LP found in reflection. To extract additional details we refer to the coupled oscillator model in Hamiltonian representation (see Methods and Fig. S5). The resulting curves, shown as green dashed lines in Fig. 4, are in good agreement with the experimental results. Such modelling allows extracting the coupling strengths of the corresponding processes (for details see SI, Fig. S5). The FP-plasmon coupling strength for the two-component system, 𝑔(2) 𝐹𝑃 ― 𝑝𝑙, turns out to be 180 meV. This value agrees well with the Rabi splitting obtained for a plasmon-cavity system and the transition dipole moment of a single plasmonic nanoparticle discussed earlier in Fig. 2. On the other hand the FP-WS2 coupling for the two-component cavity-WS2 sample, 𝑔(2) 𝐹𝑃 ― 𝑋, is only about 40 meV, significantly smaller than the corresponding cavityplasmon coupling. To better understand the three-component system, we performed FDTD calculations, which showed that plasmonic particles significantly modify the distribution of electromagnetic energy density in the cavity (see Fig. S8 for further details). To account for that within the coupled oscillator model, we introduce the plasmon-WS2 interaction constant, 𝑔(3) 𝑝𝑙 ― 𝑋, which reaches as much as 80 meV. This is enhanced in comparison to the plasmon-exciton coupling outside of the microcavity, 𝑔(2) 𝑝𝑙 ― 𝑋, which is about 50 meV (see Fig. S4). We attribute this enhancement to the antenna effect, which modifies the cavity mode such that the electromagnetic energy density is maximized at the position of WS2 monolayer due to presence of plasmonic nanoantennas in the cavity. An additional factor which supports a change of the coupling strength in this case is that here 𝑔(3) 𝑝𝑙 ― 𝑋refers just to the nanoparticle-WS2 interaction, while in Fig. S4 it is the plasmon-exciton coupling for a 13 ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 24

particle on a single mirror. It is likely that the mode volume of this reference system is larger than that for a single Au NP. The enhancement of plasmon-exciton interaction in the cavity that we observe here is in agreement with recent theoretical investigations

38, 39.

At the

(3) same time both 𝑔(3) 𝐹𝑃 ― 𝑝𝑙 and 𝑔𝐹𝑃 ― 𝑋 are nearly unaffected in the three-component system

and remain to be 180 meV and 40 meV, correspondingly. The obtained curves are in good agreement with the experiment (see green dashed lines in Fig. 4). Additionally, the coupled oscillator model allows extracting the composition of the polaritonic mixtures (see Fig. S5). In the plasmon-exciton-cavity system we observe a significant contribution of WS2 excitons for all polaritonic branches – UP, middle (MP) and LP. This is further confirmed qualitatively by FDTD calculations of absorption of the threecomponent system, individual contributions of WS2, Au nanoparticles and cavity mirrors to the total absorption (Fig. S9), as well as by their spatial and spectral distribution (Fig. S10). In experiments we, however, do not observe strong signatures of MP (except for the thickest cavity, see Fig. 3d), which we attribute to underestimated losses in the Johnson and Christy permittivity of gold 41, as mentioned in the discussion of Fig. 3. We note that the overall PL emission behavior, FDTD results (see Fig. S8-10), and Hopfield coefficients extracted from the coupled oscillator model strongly point to involvement of WS2 excitons into collective mode splitting and formation of macroscopic coherent microcavity-plasmon-exciton polaritonic states. In conclusion, we have shown that by introducing plasmonic nanoparticles inside microcavities, the collective strong coupling regime with monolayer WS2 system can be readily realized. The main mechanism responsible for these observations is the plasmonic antenna effect and large oscillator strength of plasmonic nanodisks. We anticipate that the concept of hierarchical microcavity-plasmon-exciton polaritons introduced here may lead to routine observation of collective strong coupling in microcavities without requiring high exciton densities, albeit at the expense of the reduced exciton character of polaritonic states. Several recent observations suggest that it is collective, not individual Rabi splitting, that is important for a number of physical and chemical processes, such as exciton transport and photochemistry

47-51

. By intermixing plasmons and excitons in a common microcavity

system one could therefore expect a substantial change in material properties of these strongly coupled emitters, which thereby may lead to new and exciting applications.

14 ACS Paragon Plus Environment

Page 15 of 24 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Methods: Fabrication of FP microcavities containing Au nanoparticle arrays and WS2 monolayers: 170 µm thick glass coverslips (Deckglaser #1) were cleaned in hot acetone, hot isopropanol and Piranha (H2O2:H2SO4 1:3) solution. Bottom microcavity mirror composed of 40 nm Au layer was evaporated using an e-beam evaporator (Kurt J. Lesker PVD225). SiO2 for half-cavities with various thicknesses were then deposited using STS PECVD (PlasmaEnhanced Chemical Vapor Deposition) at 300°C. In order to spatially locate monolayer flakes and align the sample for electron beam lithography Cr/Au (5 nm/25 nm) markers were patterned on SiO2 using standard UV-lithography process. Large monolayer flakes of WS2 were mechanically exfoliated from bulk crystals (HQ Graphene) on thin PDMS stamps which were later transferred on SiO2

40.

Monolayer flakes were characterized by PL, optical

contrast and Raman scattering (see Fig. S3). Square arrays of Au nanoparticle disks (height 20 nm) with various diameters and pitches were fabricated on the top of WS2 monolayer using standard e-beam lithography. In order to improve the crystallinity of the nanoparticles and remove resist residues, the samples were annealed in an inert atmosphere (Ar/4%H2) for 30 min at 300°C. PMMA layer with same thickness as the bottom SiO2 half cavity was spin coated on top of the Au nanoparticle array, followed by baking at 180°C for 5 min. Samples were completed by depositing top microcavity mirror (Au 40 nm) by e-beam evaporation. Optical Measurements: Near-normal reflection spectra were collected using a 20 objective (Nikon, NA=0.45), directed to a fiber-coupled spectrometer and normalized with reflection from a standard dielectric silver mirror. Dispersion relations in reflection were obtained by single shot imaging in the back focal imaging setup. Images were scanned using a liquid crystal filter (LCF) combined with an EM-CCD camera (Andor, iXon). For photoluminescence experiments, the sample was excited by a continuous wave 532 nm laser with in-coupled power in the range of 10-12 mW. PL signal was collected using a 40 objective (Nikon, NA=0.95) and directed to a fiber-coupled 30 cm spectrometer (Andor Shamrock SR-303i) equipped with a CCD detector (Andor iDus 420). Coupled Oscillator Model: To simulate normal modes of the coupled system, we utilize a standard coupled oscillator Hamiltonian in a 33 matrix representation. We then solve the eigen value problem. The matrix parameters consists of resonance energies and dissipation rates of all constituent sub-systems (FP cavity - 𝜔𝐹𝑃, 𝛾𝐹𝑃, plasmonic nanoparticles 15 ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 24

- 𝜔𝑝𝑙, 𝛾𝑝𝑙, and excitons - 𝜔𝑋, 𝛾𝑋) together with the different coupling strengths between (3) (3) them (𝑔(3) 𝐹𝑃 ― 𝑝𝑙, 𝑔𝐹𝑃 ― 𝑋, 𝑔𝑝𝑙 ― 𝑋). The matrix reads:

(

𝛾𝐹𝑃

𝜔𝐹𝑃(𝜃) ― 𝑖

𝐻=ℏ

𝑔(3) 𝐹𝑃 ― 𝑝𝑙 𝑔(3) 𝐹𝑃 ― 𝑋

2

𝑔(3) 𝐹𝑃 ― 𝑝𝑙

𝑔(3) 𝐹𝑃 ― 𝑋

𝛾𝑝𝑙

𝑔(3) 𝑝𝑙 ― 𝑋

𝜔𝑝𝑙 ― 𝑖 2

)

(2)

𝛾𝑋

𝑔(3) 𝑝𝑙 ― 𝑋

𝜔𝑋 ― 𝑖 2

First, we extract the resonance energy and dissipation of a FP cavity as a function of angle. The experimental reflectivity values were fitted using the expression: 𝜔𝐹𝑃 (𝜃) = 𝜔𝐹𝑃(0) +𝛼sin2 𝜃, with sin 𝜃 ∈ [ ―0.95 0.95]. Thereafter, we model the twocomponent systems using a 22 Hamiltonian matrix representation. Finally the 33 Hamiltonian (2) was used to fit experimental data of the three-component system (see Fig. 4j-l). The extracted energies and Hopfield coefficients are shown in Fig. 4 and Fig. S5 correspondingly. Finite-Difference Time-Domain (FDTD) Calculations: Numerical calculations were performed using a commercial FDTD solver: Lumerical, Inc. To obtain the reflection spectra in Fig. 1c, a 150 nm thick molecular layer was assumed to possess a Lorentzian-like dielectric function of the following form: 𝑓0𝜔20

(3)

𝜀𝑡𝑜𝑡𝑎𝑙(𝑓) = 𝜀∞ + 𝜔2 ― 2𝑖𝛾 𝜔 ― 𝜔2 0

0

where the Lorentz parameters have a standard meaning and read: 𝜀∞=2.1, 𝑓0=0.05, 𝜔0 =2.08 eV, 𝛾0=50 meV. Such material was then sandwiched between two 40 nm thick Au mirrors, with dielectric function of gold taken from Johnson and Christy

41.

To obtain the

reflection spectrum in Fig. 1d, a periodic array of Au disks (d=46 nm, h=20 nm and Λ =200 nm) was placed at the center of the microcavity, while the oscillator strength of the molecular Lorentzian layer was reduced to 𝑓0=0.006. In Figs. 3b,e the thickness of both mirrors was set to 40 nm, while the thickness of the dielectric (SiO2) layer was varied depending on the spectral requirement of first order cavity resonance. The dielectric function for Au was again taken from Johnson and Christy 41, while SiO2 was modelled as a nearly dispersion free and lossless dielectric with a refractive index of 1.45-1.47 (400 nm to 900 nm). The permittivity of monolayer WS2 was obtained 16 ACS Paragon Plus Environment

Page 17 of 24 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

from the literature 11 and used without further modifications. Contributions from all relevant excitonic components were taken into consideration. The WS2 monolayer (thickness 7 Å) was placed at the center of the microcavity. Au nanodisks (d=70 nm and h=20 nm) were placed on top of WS2. Fine meshing was used for monolayer WS2 (0.1 nm), Au disks (1 nm) and the microcavity (1 nm along z-axis) for accurate calculations.

The authors declare no competing financial interests.

Supporting Information available: See supporting information for bare Au nanodisks dark field, bright field, sample image, WS2 Raman, Hopfield Coefficients, FDTD near-field profiles, spatially and spectrally resolved absorption of the coupled system.

References: 1. 90. 2. 3.

Khitrova, G.; Gibbs, H. M.; Kira, M.; Koch, S. W.; Scherer, A. Nat Phys 2006, 2, (2), 81Törmä, P.; Barnes, W. L. Reports on Progress in Physics 2015, 78, (1), 013901. Purcell, E. M. Physical Review 1946, 69, 37-38

4. Baranov, D. G.; Wersäll, M.; Cuadra, J.; Antosiewicz, T. J.; Shegai, T. ACS Photonics 2017. 5. Kasprzak, J.; Richard, M.; Kundermann, S.; Baas, A.; Jeambrun, P.; Keeling, J. M. J.; Marchetti, F. M.; Szymanska, M. H.; Andre, R.; Staehli, J. L.; Savona, V.; Littlewood, P. B.; Deveaud, B.; Dang, L. S. Nature 2006, 443, (7110), 409-414. 6. Lidzey, D. G.; Bradley, D. D. C.; Skolnick, M. S.; Virgili, T.; Walker, S.; Whittaker, D. M. Nature 1998, 395, (6697), 53-55. 7. Hobson, P. A.; Barnes, W. L.; Lidzey, D. G.; Gehring, G. A.; Whittaker, D. M.; Skolnick, M. S.; Walker, S. Applied Physics Letters 2002, 81, (19), 3519-3521. 8. Schwartz, T.; Hutchison, J. A.; Genet, C.; Ebbesen, T. W. Physical Review Letters 2011, 106, (19), 196405. 9. Mak, K. F.; Lee, C.; Hone, J.; Shan, J.; Heinz, T. F. Physical Review Letters 2010, 105, (13), 136805. 10. Wang, Q. H.; Kalantar-Zadeh, K.; Kis, A.; Coleman, J. N.; Strano, M. S. Nat Nano 2012, 7, (11), 699-712. 11. Li, Y.; Chernikov, A.; Zhang, X.; Rigosi, A.; Hill, H. M.; van der Zande, A. M.; Chenet, D. A.; Shih, E.-M.; Hone, J.; Heinz, T. F. Physical Review B 2014, 90, (20), 205422. 12. Cuadra, J.; Baranov, D. G.; Wersäll, M.; Verre, R.; Antosiewicz, T. J.; Shegai, T. Nano Letters 2018. 17 ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 24

13. Wen, J.; Wang, H.; Wang, W.; Deng, Z.; Zhuang, C.; Zhang, Y.; Liu, F.; She, J.; Chen, J.; Chen, H.; Deng, S.; Xu, N. Nano Letters 2017, 17, (8), 4689-4697. 14. Zheng, D.; Zhang, S.; Deng, Q.; Kang, M.; Nordlander, P.; Xu, H. Nano Letters 2017, 17, (6), 3809-3814. 15. Kleemann, M.-E.; Chikkaraddy, R.; Alexeev, E. M.; Kos, D.; Carnegie, C.; Deacon, W.; de Pury, A. C.; Große, C.; de Nijs, B.; Mertens, J.; Tartakovskii, A. I.; Baumberg, J. J. Nature Communications 2017, 8, (1), 1296. 16. Liu, W.; Lee, B.; Naylor, C. H.; Ee, H.-S.; Park, J.; Johnson, A. T. C.; Agarwal, R. Nano Letters 2016, 16, (2), 1262-1269. 17. Lee, B.; Liu, W.; Naylor, C. H.; Park, J.; Malek, S. C.; Berger, J. S.; Johnson, A. T. C.; Agarwal, R. Nano Letters 2017, 17, (7), 4541-4547. 18. Liu, X.; Galfsky, T.; Sun, Z.; Xia, F.; Lin, E.-c.; Lee, Y.-H.; Kéna-Cohen, S.; Menon, V. M. Nat Photon 2015, 9, (1), 30-34. 19. Dufferwiel, S.; Schwarz, S.; Withers, F.; Trichet, A. A. P.; Li, F.; Sich, M.; Del PozoZamudio, O.; Clark, C.; Nalitov, A.; Solnyshkov, D. D.; Malpuech, G.; Novoselov, K. S.; Smith, J. M.; Skolnick, M. S.; Krizhanovskii, D. N.; Tartakovskii, A. I. Nature Communications 2015, 6, 8579. 20. Wang, S.; Li, S.; Chervy, T.; Shalabney, A.; Azzini, S.; Orgiu, E.; Hutchison, J. A.; Genet, C.; Samorì, P.; Ebbesen, T. W. Nano Letters 2016, 16, (7), 4368-4374. 21. Antosiewicz, T. J.; Apell, S. P.; Shegai, T. ACS Photonics 2014, 1, (5), 454-463. 22. Christopoulos, S.; von Högersthal, G. B. H.; Grundy, A. J. D.; Lagoudakis, P. G.; Kavokin, A. V.; Baumberg, J. J.; Christmann, G.; Butté, R.; Feltin, E.; Carlin, J. F.; Grandjean, N. Physical Review Letters 2007, 98, (12), 126405. 23. Rodriguez, S. R. K.; Rivas, J. G. Optics Express 2013, 21, (22), 27411-27421. 24. Väkeväinen, A. I.; Moerland, R. J.; Rekola, H. T.; Eskelinen, A. P.; Martikainen, J. P.; Kim, D. H.; Törmä, P. Nano Letters 2014, 14, (4), 1721-1727. 25. Wang, S.; Le-Van, Q.; Vaianella, F.; Maes, B.; Eizagirre Barker, S.; Godiksen, R. H.; Curto, A. G.; Gomez Rivas, J. arXiv preprint 2018, arXiv:1808.08388. 26. Tavis, M.; Cummings, F. W. Physical Review 1968, 170, (2), 379-384. 27. Zengin, G.; Gschneidtner, T.; Verre, R.; Shao, L.; Antosiewicz, T. J.; Moth-Poulsen, K.; Käll, M.; Shegai, T. The Journal of Physical Chemistry C 2016, 120, (37), 20588-20596. 28. Agranovich, V.; Benisty, H.; Weisbuch, C. Solid State Communications 1997, 102, (8), 631-636. 29. Andrew, P.; Barnes, W. L. Science 2000, 290, (5492), 785-788. 30. Coles, D. M.; Somaschi, N.; Michetti, P.; Clark, C.; Lagoudakis, P. G.; Savvidis, P. G.; Lidzey, D. G. Nature Materials 2014, 13, 712. 31. Zhong, X.; Chervy, T.; Wang, S.; George, J.; Thomas, A.; Hutchison, J. A.; Devaux, E.; Genet, C.; Ebbesen, T. W. Angewandte Chemie International Edition 2016, 55, (21), 6202-6206. 32. Flatten, L. C.; Coles, D. M.; He, Z.; Lidzey, D. G.; Taylor, R. A.; Warner, J. H.; Smith, J. M. Nature Communications 2017, 8, 14097. 33. Waldherr, M.; Lundt, N.; Klaas, M.; Betzold, S.; Wurdack, M.; Baumann, V.; Estrecho, E.; Nalitov, A.; Cherotchenko, E.; Cai, H.; Ostrovskaya, E. A.; Kavokin, A. V.; Tongay, S.; Klembt, S.; Höfling, S.; Schneider, C. Nature Communications 2018, 9, (1), 3286. 34. Mitra, A.; Harutyunyan, H.; Palomba, S.; Novotny, L. Optics Letters 2010, 35, (7), 953955. 35. Konrad, A.; Kern, A. M.; Brecht, M.; Meixner, A. J. Nano Letters 2015, 15, (7), 44234428. 36. Ameling, R.; Giessen, H. Nano Letters 2010, 10, (11), 4394-4398. 37. Balci, S.; Kocabas, C. Optics Letters 2015, 40, (14), 3424-3427. 38. Peng, P.; Liu, Y.-C.; Xu, D.; Cao, Q.-T.; Lu, G.; Gong, Q.; Xiao, Y.-F. Physical Review Letters 2017, 119, (23), 233901. 39. Gurlek, B.; Sandoghdar, V.; Martín-Cano, D. ACS Photonics 2018, 5, (2), 456-461. 18 ACS Paragon Plus Environment

Page 19 of 24 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

40. Andres, C.-G.; Michele, B.; Rianda, M.; Vibhor, S.; Laurens, J.; Herre, S. J. v. d. Z.; Gary, A. S. 2D Materials 2014, 1, (1), 011002. 41. Johnson, P. B.; Christy, R. W. Physical Review B 1972, 6, (12), 4370-4379. 42. Litinskaya, M.; Reineker, P.; Agranovich, V. M. Journal of Luminescence 2004, 110, (4), 364-372. 43. Wersäll, M.; Cuadra, J.; Antosiewicz, T. J.; Balci, S.; Shegai, T. Nano Letters 2017, 17, (1), 551-558. 44. Neuman, T.; Aizpurua, J. Optica 2018, 5, (10), 1247-1255. 45. Savona, V.; Andreani, L. C.; Schwendimann, P.; Quattropani, A. Solid State Communications 1995, 93, (9), 733-739. 46. Agranovich, V. M.; Litinskaia, M.; Lidzey, D. G. Physical Review B 2003, 67, (8), 085311. 47. Hutchison, J. A.; Schwartz, T.; Genet, C.; Devaux, E.; Ebbesen, T. W. Angewandte Chemie International Edition 2012, 51, (7), 1592-1596. 48. Feist, J.; Garcia-Vidal, F. J. Physical Review Letters 2015, 114, (19), 196402. 49. Herrera, F.; Spano, F. C. Physical Review Letters 2016, 116, (23), 238301. 50. Galego, J.; Garcia-Vidal, F. J.; Feist, J. Nature Communications 2016, 7, 13841. 51. Munkhbat, B.; Wersäll, M.; Baranov, D. G.; Antosiewicz, T. J.; Shegai, T. Science Advances 2018, 4, (7), eaas9552.

19 ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 24

TOC Figure

20 ACS Paragon Plus Environment

Page 21 of 24

a

b

͌

1.0

1.0

c

d 0.8

R/R0

0.8

R/R0

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

0.6

0.6 ~280meV

0.4

0.4

~280meV 1.5

2.0

2.5

3.0

1.5

Energy (eV)

ACS Paragon Plus Environment

2.0

2.5

Energy (eV)

3.0

Nano Letters

L, nm 250

200

a

c

g, meV

b

300

280 240

d

L

200

˄

160 Cavity thickness 120 1.0

1.0

R (exp.)

0.8

160nm 180nm 200nm

1.0

e

d

f 0.8

0.8

0.6 0.6 0.4

0.6 W≈500meV

empty cavity d=80nm, L=200nm d=80nm, L=250nm d=80nm, L=300nm

L=160 nm

1.0

R (calc.)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 24

g

L=180 nm 0.4 1.0

h

L=200 nm 0.4 1.0

0.8

0.8

0.8

0.6

0.6

0.6

0.4

0.4

0.4

0.2 1.4

L=160 nm 1.6

1.8

2.0

2.2

Energy, eV

2.4

0.2 1.4

L=180 nm 1.6

1.8

2.0

2.2

2.4

Energy, eV

ACS Paragon Plus Environment

i

0.2 1.4

L=200 nm 1.6

1.8

2.0

2.2

Energy, eV

2.4

Page 23 of 24

cavity + WS2 180nm

R (exp.)

170nm

LP

UP

cavity + WS2 + NPs

d

a LP

UP

160nm 150nm 140nm

LP

b

LP

UP

e

R (calc.)

UP

LP X

c

f

LP

PL (exp.)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

X

1.4 1.6

1.8 2.0 2.2

2.4 1.4 1.6

Energy (eV)

ACS Paragon Plus Environment

1.8 2.0 2.2

Energy (eV)

2.4

Nano Letters

140 nm

180 nm

160 nm

q

Energy (eV)

2.4 2.2 2 1.8 1.6

a

b

Energy (eV)

2.4

R

empty cavity

c

PL

R

PL

R

PL

2.2 2 1.8 1.6

d

e

f

cavity+WS2

g

h

i

cavity+NPs

Energy (eV)

2.4 2.2 2 1.8 1.6 2.4 Energy (eV)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 24

R

PL

R

PL

R

PL

2.2 2 1.8 1.6

j

k -0.6

0 sinq

0.6

cavity+WS2+NPs

l -0.6

0 sinq

0.6

-0.6

ACS Paragon Plus Environment

0 sinq

0.6